> > On 29 August 2018 at 06:47 "Janusz S. Bień via Unicode" > <unicode@unicode.org> wrote: > > > > > > Storing this information in a font, by hook or crook, would lock > > users > > of those PUA characters into that font. At that rate, you might as > > well > > use ASCII-hacked fonts, as we did 25 years ago. > > > > >
I don't see that at all. The obvious way in the sfnt format, used by OpenType, is as a table consisting entirely of the XML file. It is quite easy to add a table to an unsigned sfnt font, and even easier to extract a table consisting entirely of UTF-8 text, though ASCII would be even easier, from a font file. > > Storing the information in a font is inappropriate not only for > thetechnical reasons, as I wrote recently (on Thu, Aug 23 2018) > > > > > > Fonts are for *rendering*, new characters and variants are more and > > more often needed for *input* of real life old texts with sufficient > > precision. > > > > > 1. There are existing methods of associating a font with a text. Not using a font needs a new scheme for associating a set of PUA properties with a portion of a file. The font also serves as a code chart. It can also hold information on how characters combine, which is notoriously beyond the capability of code charts. 2. Registries can vanish. 3. In practice, a file needs to retain an association with a specialist font. Preserving the font should preserve its content, but there are pruning techniques (e.g. WOFF2) that may remove this content. Richard.